S1PR1 regulates lymphatic valve development and tertiary lymphoid organ formation in the ileum

Abstract

Efficient lymph flow is ensured by lymphatic valves (LVs). The mechanisms that regulate LV development are incompletely understood. Here, we show that the deletion of the GPCR sphingosine 1-phosphate receptor-1 (S1PR1) from lymphatic endothelial cells (LECs) results in fewer LVs. Interestingly, LVs that remained in the terminal ileum-draining lymphatic vessels were specifically dysfunctional. Furthermore, tertiary lymphoid organs (TLOs) formed in the terminal ileum of the mutant mice. TLOs in this location are associated with ileitis in humans and mice. However, mice lacking S1PR1 did not develop obvious characteristics of ileitis. Mechanistically, S1PR1 regulates shear stress signaling and the expression of the valve-regulatory molecules FOXC2 and connexin-37. Importantly, Foxc2+/- mice, a model for lymphedema-distichiasis syndrome, also develop TLOs in the terminal ileum. Thus, we have discovered S1PR1 as a previously unknown regulator of LV and TLO development. We also suggest that TLOs are a sign of subclinical inflammation that can form due to lymphatic disorders in the absence of ileitis.

Graphical abstract

Figure 1.

S1PR1 signaling is active during and is necessary for LVV and venous valve morphogenesis. (A) E16.5 S1PR1-GS and H2B-GFP (control) littermates were frontally sectioned and analyzed. GFP⁺ cells were observed in the LVVs (arrows) and venous valves (arrowheads) of S1PR1-GS embryos, indicating that S1PR1 signaling is active in these structures. Only a few GFP⁺ cells were observed in H2B-GFP embryos due to leaky expression of this transgene. (B) 500-μm–thick transverse section of E16.5 Prox1-tdTomato and Lyve1-Cre;S1pr1^−/f; Prox1-tdTomato embryos were prepared using a vibratome, and whole-mount IHC and confocal imaging were performed to visualize the LVVs and venous valves (bottom row). Subsequently, the same samples were processed and analyzed by SEM (top row). Normal-looking LVVs (arrows or pseudo colored in magenta) and venous valves (arrowheads or pseudo colored in green) were seen in control embryos. Mutant embryos appeared to have substantially fewer valvular endothelial cells. When present the mutant cells appeared disorganized and arrested at the periphery of the vessels. (C and D) E16.5 control, Lyve1-Cre;S1pr1^−/f, and Lyve1-Cre;S1pr1^−/f;Vegfr3^+/− embryos were frontally sectioned (12-μm thick) and analyzed by IHC. LVVs (arrows) and venous valves (arrowheads) of control embryos had invaginated into the veins. Higher magnification images of control LVVs (boxed areas and the panels below in C) revealed that VEGFR3^hi;Prox1^hi LECs (asterisks) were located in between two layers of VEGFR3^Low;Prox1^hi valvular endothelial cells (yellow arrowheads). In contrast, invagination was substantially reduced in the LVVs and venous valves of mutant embryos, and the organization of the two cell types was defective. Blood cells were observed within the lymph sacs of mutant embryos (yellow arrows in C). (D) The invagination defect of Lyve1-Cre;S1pr1^−/f embryos was not rescued by Vegfr3 heterozygosity. Statistics: (A, C, and D) n = 5 embryos per genotype; (B) n = 5 controls and n = 7 Lyve1-Cre;S1pr1^−/f LVV complexes. LS, lymph sacs; T, thymus; EJV, external jugular vein; SVC, superior vena cava; SCV, subclavian vein; IJV, internal jugular vein.

Figure S1.

S1PR1 regulates dermal LV development in a VEGFR3-independent manner. The dorsal skin of E16.5 embryos was dissected and analyzed by whole-mount IHC with anti-PROX1 antibody. The developing LVs were identified by the presence of PROX1^hi clusters (arrowheads). Statistics: n = 5 for control embryos; n = 4 for each mutant genotype, respectively. Each dot represents one animal on the graph. The graphs were plotted as mean ± SD. One-way ANOVA was performed for the statistical analysis. ***P < 0.001; ****P < 0.0001.

Figure 2.

S1PR1 signaling is active in LVs and is necessary for LV development. (A) The mesenteries of P10 S1PR1-GFP and H2B-GFP littermates where analyzed. Collecting lymphatic vessels of S1PR1-GFP pups were GFP⁺ and GFP expression was stronger in the LVs (arrowheads). GFP expression was not observed in H2B-GFP pups. (B) The mesenteries of P10 S1pr1^f/f;Prox1-tdTomato and S1pr1^iΔLEC;Prox1-tdTomato pups that were administered TM from P1–7 were analyzed. Mesenteric tissue that was connected to the duodenum and jejunum was considered “proximal,” and that which is connected to the ileum and cecum was considered “distal,” LVs (puncta) in the proximal and distal vessels were counted and quantified. LVs were significantly reduced in S1pr1^iΔLEC;Prox1-tdTomato pups. The reduction appeared to be more severe in the posterior section of the gut. (C) The mesenteric lymphatic vessels and LVs of P10 pups that were administered TM from P1–7 were analyzed using the indicated antibodies. CX37 expression appeared to be reduced in the remaining LVs of S1pr1^iΔLEC pups. Statistics: (A) n = 9 S1PR1-GS and n = 3 H2B-GFP pups; (B) n = 5 control and n = 8 S1pr1^iΔLEC pups. Each dot represents an individual animal on the graph for the total number of LVs. Three proximal vessels and three distal vessels from each mesentery were analyzed to quantify the valve density. Each dot represents a vessel on the graph; (C) n = 3 per genotype. The graphs were plotted as mean ± SD. Unpaired t test was performed for the statistical analysis. ****P < 0.0001.

Figure S2.

Tg(Prox1-CreERT2) efficiently targets the entire mesenteric lymphatic vasculature and depletes S1PR1. (A) The proximal and distal mesenteric vasculature of P10 Tg(Prox1-CreERT2);R26^mT/mG (TM@P1–7) pups were analyzed by IHC for PROX1 and autofluorescence of membrane-targeted GFP and membrane-targeted tdTomato. Downregulation of tdTomato and upregulation of GFP confirmed the efficient targeting of proximal and distal mesenteric lymphatic vessels by Tg(Prox1-CreERT2). GFP signal was semiquantitatively measured and plotted. (B) The proximal and distal mesenteric vasculature of P10 S1pr1^f/f (TM@P1–7) and S1pr1^iΔLEC (TM@P1–7) littermates were analyzed using the indicated antibodies. Semiquantitative measurement of the fluorescence signal confirmed that S1PR1 was specifically and efficiently deleted from the lymphatic vessels of S1pr1^iΔLEC pups. Arrows point to blood capillaries in which S1PR1 expression was comparable between the two genotypes. RFA, relative fluorescence activity. Scale bar is 100 μm. Statistics: (A) n = 4. Each dot represents the fluorescent intensity of a single vessel. Three proximal and three distal vessels from each mesentery were measured. (B)n = 4 S1pr1^f/f; n = 3 S1pr1^iΔLEC pups. Each dot represents the RFA of a single vessel. Three proximal and three distal vessels from each mesentery were measured. The graphs were plotted as mean ± SD. Unpaired t test (A) and two-way ANOVA (B) were performed for the statistical analysis. ***P < 0.001; ****P < 0.0001.

Figure S3.

Remaining LVs after efficient deletion of S1PR1 express valve markers normally. The mesenteric vasculature of P10 S1pr1^f/f (TM@P1–7) and S1pr1^iΔLEC (TM@P1–7) littermates was analyzed using the indicated antibodies. (A and B) The expression of valve markers PROX1 (A and B), GATA2 (B), and integrin-α9 (A) appeared normal in the remaining LVs of mutants. Statistics: n = 3 pups per genotype.

Figure 3.

LVs are reduced in the numbers and defective in S1pr1 ^iΔLEC mice. (A and B) 3-mo-old S1pr1^iΔLEC mice that were treated with TM either from P1–P7 (A) or at 8 wk of age (B) were studied. (A) Dermal lymphatic vessels in the ears of S1pr1^iΔLEC (TM@P1–P7) mice had more branches and fewer claudin-5^hi LVs (arrowheads) when compared with control littermates. (B)S1pr1^iΔLEC (TM@8w) mice had elevated number of branch points but did not have any obvious reduction in LVs. (C) Representative stitched image of a terminal ileum-draining mesenteric lymphatic vessel of a control mouse with LVs (arrows) is shown. A corresponding stitched image of a lymphatic vessel from an S1pr1^iΔLEC (TM@P1–7) mouse lacking LVs is also shown (red arrowhead indicates where small nubs remain from a valve). The graph shows that the LV density is significantly reduced in S1pr1^iΔLEC (TM@P1–7) but not in S1pr1^iΔLEC (TM@8w) mice. (D) Ex vivo analysis of LVs in the ileum-draining lymphatic vessels. LVs of control, S1pr1^iΔLEC (TM@P1–7), and S1pr1^iΔLEC (TM@8w) mice were analyzed for back leak. The graph shows that the LVs in the terminal ileum-draining lymphatic vessels of mutant mice were significantly leaky irrespective of the time of gene deletion. (E) Following the back leak analysis, some leaky valves fixed for subsequent immunofluorescence to assess the specific LV structural component(s). IHC was performed on isolated vessel for the indicated markers, imaged by confocal microscopy, and 3D reconstructed. The 3D images were rotated to visualize the LVs on their side (left) or en face (right). LVs with two symmetrical leaflets were observed in control mice. However, in 1 of 6 S1pr1^iΔLEC (TM@P1–7) LVs imaged, only one partial leaflet (arrows, Prox1^Hi-ITGA9⁺) was observed at the LV site, which resulted in complete back leak. (F) We performed live confocal imaging on 3 S1pr1^iΔLEC (TM@P1–7) LVs that exhibited various levels of back leak and two control LVs without back leak using the Prox1-tdTomato reporter under various levels of Pin and Pout. Both control and mutant LVs remained open when Pin and Pout were equal and closed when Pout was slightly elevated. The control LV remained closed when Pout was increased to 8 cm H₂O. In contrast, a gap remained in a mutant LV (arrow), resulting in back leak. Symmetrically located commissures that extend in the downstream direction can be observed in the same control LV visualized from the side (yellow arrows). In contrast, the S1pr1^iΔLEC (TM@P1–7) LV had asymmetrical commissures that extended in both upstream and downstream directions (yellow arrows). Statistics: (A and B) Each dot represents an individual mouse. The graphs were plotted as mean ± SD. Mann–Whitney test and unpaired t test were performed for the statistical analysis. *P < 0.05. (C) LV density was measured in ileum-draining lymphatic vessels harvested from n = 10 3-mo S1pr1^f/f (TM@P1–P7), n = 5 10-mo S1pr1^f/f (TM@P1–P7), n = 11 3-mo S1pr1^iΔLEC (TM@P1–P7), and n = 3 10-mo S1pr1^iΔLEC (TM@8w) mice. One-way ANOVA with Tukey’s post hoc test was performed to determine significance. ***P < 0.001. (D) Each dot represents an individual LV harvested from n = 10 3-mo S1pr1^f/f (TM@P1–P7), n = 2 10-mo S1pr1^f/f (TM@P1–P7), n = 11 3-mo S1pr1^iΔLEC (TM@P1–P7), and n = 3 10-mo S1pr1^iΔLEC (TM@8w) mice. A nonparametric Kruskal–Wallis test with Dunn’s post hoc test was performed to determine significance. *P < 0.05; **P < 0.01. (E and F) Live imaging followed by fixation and whole-mount IHC was performed using n = 2 LVs from 3-mo S1pr1^f/f;Prox1-tdTomato (TM@P1–P7) and n = 3 LVs from 3-mo S1pr1^iΔLEC;Prox1-tdTomato (TM@P1–P7) mice. Additionally, n = 3 LVs that were not imaged live from 3-mo S1pr1^iΔLEC (TM@P1–P7) mice were directly fixed and imaged by whole-mount IHC.

Figure S4.

Most LVs located in the proximal mesenteric lymphatic vessels of S1pr1 ^iΔLEC mice were normal. (A) Mesenteric lymphatic vessels from the sections corresponding to the duodenum, jejunum, and ileum were dissected from 3-mo-old S1pr1^iΔLEC (TM@P1–7) mice, cleaned, and cannulated (top row). Subsequently, LV function test was performed. While gradually increasing the Pout, the pressure (Psn) and vessel diameter were measured behind the LVs. Duodenal and jejunal LVs did not have any pressure back leak (second row). Consequently, those vessels did not expand in diameter as Pout was raised (bottom row). In contrast, an LV from the ileum exhibited back leak, as indicated by increasing Psn and diameter with increasing Pout. (B) Quantification of back leak of LVs harvested from the duodenum and jejunum versus ileum of control and S1pr1^iΔLEC mice. Statistics: (B) n = 31 control (duodenum and jejunum), n = 28 control (ileum), n = 15 S1pr1^iΔLEC (duodenum and jejunum), and n = 26 S1pr1^iΔLEC (ileum) LVs. LVs from 3-mo-old S1pr1^iΔLEC (TM@P1–P7) and 10-mo-old S1pr1^iΔLEC (TM@8w) were combined as S1pr1^iΔLEC for this analysis. *P < 0.05; **P < 0.01.

Figure S5.

Heterogeneous LV defect in S1pr1 ^iΔLEC mice. (A) Lymphatic vessels with leaky LVs from 3-mo-old S1pr1^iΔLEC (TM@P1–7) mice were fixed, and whole-mount IHC was performed for the indicated markers. Confocal imaging and 3D reconstruction were performed to identify the structural defects in the LVs. One LV appeared to have three leaflets (arrows). One or both leaflets were abnormally elongated at their insertion points in two other LVs, resulting in abnormal en face LV structure (right). (B) Back leak test revealed significant leakage in the ileal LVs of 4-mo-old S1pr1^iΔLEC (TM@8w). (C) A leaky LV from a 4-mo-old S1pr1^iΔLEC (TM@8w) mouse was fixed, and whole-mount IHC was performed for PROX1 (blue) and CD31 (red). Confocal imaging and 3D reconstruction revealed a single leaflet (arrow). (D) Lymphatic vessels were incubated with CellTracker Green (CMFDA) stain and imaged live under various Pins and Pouts. The confocal images were 3D reconstructed or analyzed at various planes. Control LV was open when Pin = Pout and closed when Pout > Pin. Tight overlap between leaflets can be observed in digital sections. In contrast, a leaky LV from a 4-mo-old S1pr1^iΔLEC (TM@8w) mouse remained open when Pout > Pin. Digital sections revealed short leaflets that did not overlap. Statistics: (A) Live imaging followed by fixation and whole-mount IHC was performed using n = 2 LVs from 3-mo-old S1pr1^f/f;Prox1-tdTomato (TM@P1–P7) and n = 3 LVs from 3-mo-old S1pr1^iΔLEC;Prox1-tdTomato (TM@P1–P7) mice. Additionally, n = 3 LVs that were not imaged live from 3-mo-old S1pr1^iΔLEC (TM@P1–P7) mice were directly fixed and imaged by whole-mount IHC. (B) n = 13 LVs from n = 8 4-mo-old S1pr1^f/f (TM@8w) mice and n = 13 LVs from n = 5 4-mo-old S1pr1^iΔLEC (TM@8w) mice were analyzed for back leak. **P < 0.01. (C and D) Following back leak test (B), n = 5 LVs from 4-mo-old S1pr1^f/f (TM@8w) mice and n = 8 leaky LVs from 4-mo-old S1pr1^iΔLEC (TM@8w) mice were analyzed by live imaging, followed by fixation and whole-mount IHC. In addition to those 8, the single leaflet valve from the S1pr1^iΔLEC (TM@8w) mouse was assessed only by fixation and whole-mount IHC (C). (D) Under live imaging at high pressure, three control valves from 4-mo-old S1pr1^f/f (TM@8w) mice were completely normal. A representative control valve is shown in D. One control valve had a small gap at the commissure at the high pressure. The remaining valve had 1 dysfunctional commissure where the annulus failed to meet and asymmetrical leaflet insertions into the sinus, which prevented it from closing in response to adverse pressure. Of the S1pr1^iΔLEC (TM@8w) LVs at high pressures, four LVs failed to close while three were partly closed and one tightly closed. From the four leaky LVs that failed to close, one had short leaflets that did not reach the midpoint of the lumen (shown in D). In two of the four leaky LVs and two of the three partly leaky LVs, there was asymmetry in the leaflet’s upstream insertion site.

Figure 4.

Lymphatic drainage is defective and obstructed by nodules in S1pr1 ^iΔLEC mice. (A and B) FITC-conjugated dextran (2,000 kD) was injected into the muscle layer of the ileum and/or the Peyer’s patches of anesthetized 3-mo-old (A) or 10-mo-old (B) S1pr1^f/f and S1pr1^iΔLEC mice that were treated with TM from P1–7. The flow of fluorescent dye through the mesenteric lymphatic vessels was visualized by live imaging. (A) The dye rapidly drains through the lymphatic vessels in control mice. In contrast, the dye abruptly stopped (white arrow) or appeared to bypass certain locations (yellow arrows) in mutant mice. The graphs show that the distance travelled by the dye and the rate at which the dye travelled were significantly reduced in S1pr1^iΔLEC mice. (B) Time in seconds after injection is indicated on the top right corner of the panels. The dye rapidly drains through the lymphatic vessels in control mice (white arrows). In contrast, the dye accumulated in nodules that were connected to the lymphatic vessels in the mutant mice (white arrowheads). Retrograde flow was also observed between the nodules (yellow arrows). The nodules were observed both in pre-collecting vessels (red arrowheads) and in collecting lymphatic vessels (red arrow). The videos were analyzed to quantify the distance travelled by the dye and the rate of travel. The graphs show that these parameters were significantly reduced in S1pr1^iΔLEC mice. Statistics: Images are representative of (A) n = 3 S1pr1^f/f and n = 4 S1pr1^iΔLEC mice; (B) n = 4 S1pr1^f/f and n = 5 S1pr1^iΔLEC mice. Some samples were analyzed by injection at multiple sites. Each dot in the graph indicates an individual injection. Graphs were plotted as mean ± SEM. Unpaired t tests were performed for the statistical analyses. *P < 0.05; ****P < 0.0001.

Figure 5.

TLOs are present in the terminal ileum of S1pr1 ^iΔLEC mice, and they are connected to the lymphatic vessels. (A) Mesenteric tissue along the ileum was harvested and analyzed using the lymphatic vessel marker VEGFR3. Stitched images revealed a large number of nodules in S1pr1^iΔLEC (TM@P1–7) mice, but not in control littermates. The nodules were observed both in pre-collecting vessels (arrowheads) and in collecting lymphatic vessels (arrows). (B) The number and size of LYVE1⁺B220⁺ nodules were measured and quantified. (C) Mesenteries of 8–12-mo-old control mice and S1pr1^iΔLEC littermates that were treated with TM from P1–7 were analyzed using markers for the immune cells, stromal cells, and endothelial cells. Some S1pr1^iΔLEC mice had R26^+/tdTomato reporter, the expression of which was permanently induced in the lymphatic vessels by TM-activated CreERT2. A few other mice had the Prox1-tdTomato reporter. Lymphatic vessels were labelled by VEGFR3, PROX1, LYVE1, and tdTomato. LYVE1 was also expressed in a subset of macrophages. B220, CD3e, CD11b, and S100A9 are markers of B cells, T cells, myeloid lineage cells, and neutrophils, respectively. CCL21 is a marker for lymphatic vessels and the FRCs within TLOs. PECAM1 labels all endothelial cells, but its expression is stronger in blood endothelial cells when compared with LECs. ICAM1 is expressed in HEVs, inflamed blood endothelial cells, and a variety of immune cells. MAdCAM1 is a marker of HEVs. (D) TLOs in the terminal ileum of 1-year-old S1pr1^f/f and S1pr1^iΔLEC (TM@8W) mice were counted and plotted. (E) Mesenteric lymphatic vessels of S1pr1^f/f mice had clear claudin-5⁺ LVs (arrows) and weak LYVE1 expression. LVs were also observed in S1pr1^iΔLEC (TM@8W) mice (arrows), although those that were close to the TLOs appeared defective. The TLOs had GL7⁺ germinal center B cells, CD11c⁺ dendritic cells, and F4/80⁺ macrophages. Statistics: (A) n = 3 S1pr1^f/f and n = 3 S1pr1^iΔLEC (TM@P1–7) mice; (B) n = 6 S1pr1^f/f and n = 8 S1pr1^iΔLEC mice. The size of the nodules was quantified, and each dot represents a nodule on the graph; (C and E) representative images from three to five mice/genotype/marker; (D) n = 6 S1pr1^f/f and n = 7 S1pr1^iΔLEC (TM@8W) mice. Graphs were plotted as mean ± SD. Welch’s t test (B [number of TLOs] and D) and Mann–Whitney test (B [TLO size]) were performed for the statistical analysis. ***P < 0.001; ****P < 0.0001.

Figure 6.

Autocrine or paracrine S1PR1 signaling regulates LV development and TLO formation. (A and B) Mesenteric (A) and dermal (B) lymphatic vessels of Sphk1/2^ΔLEC mice, in which S1P synthesis in LECs was ablated, were analyzed. (A) Stitched images of mesenteric lymphatic vessels revealed fewer LVs in P10 Sphk1/2^ΔLEC pups. The remaining LVs appeared immature (bottom row). Representative valves are within dotted boxes, and their enlarged images are shown below. (B) Dermal lymphatic vessels of 3-mo-old Sphk1/2^ΔLEC mice had fewer LVs (yellow arrows) and more branches per field. (C) The mesenteries of 1-year-old S1PR1-GS and S1PR1-GS;Sphk1/2^ΔLEC mice were analyzed. GFP autofluorescence was observed in the lymphatic vessels of S1PR1-GS mice (arrows). Weaker GFP expression was observed in blood vessels. In contrast, lymphatic vessels could not be identified based on GFP autofluorescence in S1PR1-GS;Sphk1/2^ΔLEC mice. However, GFP expression was observed in blood vessels and in nodule-like structures (arrows). (D) IF for GFP and VEGFR3 revealed a dramatic downregulation of GFP expression in the lymphatic vessels of S1PR1-GS;Sphk1/2^ΔLEC mice. (E) IF for LYVE1 and B220 revealed significant number of TLOs in 3-, 6- and 12-mo-old Sphk1/2^ΔLEC mice. The picture shows a representative TLO from a 12-mo-old Sphk1/2^ΔLEC mouse. Statistics: (A) n = 3 control and n = 4 Sphk1/2^ΔLEC pups. Total number of LVs in the entire mesentery were counted. Each dot in the graph represents an individual animal. Three proximal vessels and three distal vessels from each mesentery were analyzed to quantify the valve density. Each dot in the graph represents a vessel; (B) n = 4 controls and n = 3 Sphk1/2^ΔLEC mice. (C) n = 7 S1PR1-GS and n = 8 S1PR1-GS;Sphk1/2^ΔLEC mice; (D) n = 3 6-mo-old S1PR1-GS and n = 3 S1PR1-GS;Sphk1/2^ΔLEC littermates; (E) 3-mo-old mice: n = 4 control and n = 4 Sphk1/2^ΔLEC; 6-mo-old mice: n = 6 control and n = 8 Sphk1/2^ΔLEC; and 1-year-old mice: n = 5 control and n = 5 Sphk1/2^ΔLEC. Graphs were plotted as mean ± SD. Unpaired t tests (A), Welch’s t tests (D), and Mann–Whitney tests (B and E) were performed for the statistical analysis. *P < 0.05; **P < 0.01; ***P < 0.001.

Figure 7.

S1pr1 ^iΔLEC mice did not have obvious characteristics of inflammatory disease. (A) 10-mo-old S1pr1^f/f (TM@P1–7) and S1pr1^iΔLEC (TM@P1–7) littermates had comparable body weights irrespective of sex. (B) The number of Peyer’s patches in the guts (duodenum to cecum) of 10-mo-old S1pr1^f/f (TM@P1–7) and S1pr1^iΔLEC (TM@P1–7) littermates was counted and found to be comparable. (C and D) Spleen (C) and mesenteric LNs (D) of 10-mo-old S1pr1^f/f (TM@P1–7) and S1pr1^iΔLEC (TM@P1–7) littermates were comparable in size and shape. Statistics: (A) Each dot in the graph represents an individual animal. n = 4 male S1pr1^f/f, n = 4 male S1pr1^iΔLEC, n = 2 female S1pr1^f/f, and n = 4 female S1pr1^iΔLEC. Statistical significance was calculated for males using Mann–Whitney test; (B) n = 4 males and 1 female per genotype. Statistical significance was calculated using Mann–Whitney test; (C and D) images are representative of n = 3 animals/genotype/sex.

Figure 8.

S1pr1 ^iΔLEC mice do not develop epithelial dysplasia, ileal inflammation, or microbial dysbiosis. (A and B) The ileum of S1pr1^f/f and S1pr1^iΔLEC mice (TM@8w) were sectioned and analyzed by H&E (A) or IHC for the indicated markers (B). (A) The intestinal epithelium of S1pr1^iΔLEC mice appeared to be indistinguishable from control samples with no obvious infiltration of immune cells. (B) The same number of CD45⁺ hematopoietic cells, B220⁺ B cells, and S100A9⁺ neutrophils were observed in the control and S1pr1^iΔLEC mice. (C) The serum from S1pr1^f/f and S1pr1^iΔLEC mice was analyzed by multiplex ELISA for inflammatory cytokines. IL-7 was modestly increased in the S1pr1^iΔLEC mice. No significant differences were observed between the control and S1pr1^iΔLEC mice for the other cytokines. (D) The fecal pellets of S1pr1^f/f and S1pr1^iΔLEC mice were analyzed by shotgun metagenomic analysis. The numbers on top of the graphs indicate the number of bacterial reads. The same information is provided within brackets as a percentage of total reads. The graphs indicate the relative abundance of the bacterial species. No obvious change was observed between the S1pr1^f/f and S1pr1^iΔLEC mice. The inset is a hierarchical clustering of the same samples with the co-housed mice within red boxes. Three of the four co-housed mice appear to be closer to each other irrespective of their genotype. Statistics: (A and B) n = 4 control and n = 5 S1pr1^iΔLEC mice (TM@8w). Two to four sections per mice were analyzed, and the number of cells per crypt-villus unit was calculated, and graphs were plotted as mean ± SEM. Statistical significance was calculated using unpaired t test for CD45 and S100A9 and Mann–Whitney test for B220; (C) n = 8 control and n = 10 S1pr1^iΔLEC mice (TM@8w). Graphs were plotted as mean ± SD. Significance was determined using unpaired t test with Welch’s correction for MCP-1 and RANTES and Mann–Whitney test for the other cytokines. (D)n = 2 male S1pr1^iΔLEC (TM@P1–7) and n = 3 TM-treated S1pr1^f/f control littermates. In the distance plot, cohoused mice are within the red box. *P < 0.05.

Figure 9.

S1PR1 regulates OSS response and the expression of valve regulatory genes in HLECs. (A) HLECs were transfected with siControl or siS1PR1 and grown for 24 h under static conditions to knockdown S1PR1. Subsequently, cells were cultured under static or OSS for 24 h. Cells were immunostained for F-actin or VE-cadherin. F-actin was primarily located along the cell wall (cortical actin) of control cells under both static and OSS conditions. Control HLECs became more spherical, and cortical actin expression appeared to be increased by OSS. In contrast, siS1PR1-transfected HLECs appeared elongated and had increased expression of stress fibers. The percentage of VE-cadherin⁺ overlapping cell junctions was increased by OSS, and this enhancement was abolished by siS1PR1. (B) HLECs were cultured as described above, and western blotting was performed for the indicated proteins. OSS induced the expression of the shear stress-responsive transcription factor KLF4 and the valve-regulatory molecules active β-catenin, FOXC2, and CX37. Knockdown of S1PR1 significantly inhibited the expression of these molecules. (C) HLECs were transfected with siControl or siS1PR1 and grown for 48 h under static conditions to knockdown S1PR1. Subsequently, cells were cultured under static or OSS for 10 min. Cell lysates were western blotted for the indicated antibodies, and quantified. pAKT, pERK, and pFOXO1 were upregulated by OSS. siS1PR1 significantly downregulated the expression of pAKT and pFOXO1. (D) Schematic summary of OSS response in LECs. S1PR1 preserves VE-cadherin and cortical actin and promotes the phosphorylation of AKT. Phosphorylated AKT promotes the phosphorylation and nuclear exclusion of FOXO1 and prevents the phosphorylation and degradation of β-catenin. The later two processes are likely responsible for the expression of valve-regulatory molecules FOXC2 and CX37 (encoded by GJA4) and the shear-stress responsive transcription factor KLF4. In the absence of S1PR1, LECs lose VE-cadherin, gain stress fibers, become elongated, and do not upregulate valve-regulatory genes or KLF4 in response to OSS. (E) The mesenteric tissues from 1-year-old control and Foxc2^+/− mice were analyzed by IHC for the indicated markers to identify and quantify TLOs. A representative TLO from a Foxc2^+/− mouse is shown. A significantly higher number of TLOs were observed in the Foxc2^+/− mice. Statistics: (A) The axis was measured in 30 cells, and the junction was analyzed in 20 cells in a single field from each of the three experiments. Each dot represents one cell in the graphs; (B and C) the blot is representative of three independent experiments. The data from all experiments were used to prepare the graphs; (E) each dot in the graph indicates an individual mouse. n = 5 controls, n = 9 Foxc2^+/− mice. The graphs are shown as mean ± SD. Two-way ANOVA with Tukey’s post hoc test (A–C) and unpaired t test with Welch’s correction (E) were performed to determine statistical significance. *P < 0.05; **P < 0.01; ***P < 0.001; ****P < 0.0001. Source data are available for this figure: SourceData F9.

Figure 10.

Working model. LVs are necessary for proper immune cell drainage to the LNs. S1PR1 regulates LV development by enhancing the expression of FOXC2 and CX37 in response to OSS. The loss of S1PR1 results in reduced lymph flow and TLO formation.